Studies in Personal Space ^

ROBERT SOMMER, The Saskatchewan Hospital

Surprisingly little is known about the way people use space. Social scientistsin the field of human ecology have been concerned primarily with the distri-bution of such things as social classes, economic institutions, and mental illness.An almost unexplored area is microecology or the way that people in pairs orsmall groups arrange themselves.

In the studies of Hediger (5), Howard (6), and 'Von Uexkull (15), "space"has had two different meanings. The more familiar of these refers to space inthe geographic sense, i.e., space as area. It is most commonly discussed withreference to the animal's territory or home. However, some writers have ap-plied the concept of territory to human behavior. W. F. Whyte(16) andThrasher (14) have studied the territories of adolescent gangs. W. H. Whyte(17) also has studied the groupings of people within a housing development.Probably the most intensive investigation thus far has been that of Barkerand Wright (2) in the Midwest. Yet this still remains relatively unexploredterritory for social scientists.

The second way in which the term "space" is used can be called "personalspace of the organism." Although it has its roots in the work of zoologists andethologists, it is an entirely different concept from that of territory. Personaldistance is the distance that the organism customarily places between itselfand other organisms. This distance may vary from species to species and indi-vidual to individual. Hediger speaks of this as "flight distance" and has meas-ured this for hundreds of animals.

This concept would seem to have relevance for the study of human behavior,although it has never been studied empirically so far as we know. It seemsobvious that people feel uncomfortable when they talk to others who eitherstand too close or too far away. The concept itself has been used several timesin the literature. David Katz (8) used the term "personal space" and comparedit to the shell of a snail. Von Uexkull used the graphic analogy of people"surrounded by soap-bubble worlds." Stern (13) developed the concept ofpersonal world. He noted that the physical world was without a center, but the

1 This study was supported by grants from the Rockefeller Foundation, the Province ofSaskatchewan, and the Department of Health and Welfare (Ottawa). Drs. HumphryOsmond ajid T. E. Weckowicz gave valuable advice and encouragement. The followingpeople were research assistants during the study: R. Dewar, R, Hall, Eleanor Fenton,Sheila Kelly, Gwen Witney, Dorothy Sommer, Sandra Wright.

247

248 SOCIOMETRY

personal world had a natural center from which and toward which everythingpertaining to it extends. This center is the person himself about whom thepersonal world is oriented. Stern's analogy to the "snail shell" of Katz and the"soap bubble" of Von Uexkull was to describe the ''personally near" as an"aura" surrounding the person.

The concepts of "personal space" can be distinguished from that of "terri-tory" in several ways. The most important difference is that personal space iscarried around while territory is relatively stationary. The animal or man willusually mark the boundaries of his territory so that they are visible to others,but the boundaries of personal space are invisible. Personal space has thebody as its center, while territory does not. Often the center of territory is thehome of the animal or man. Animals will usually fight to maintain dominionover their territory but will withdraw if others intrude into their personalspace.

This paper will describe several studies of personal distance. The first studywas purely observational. That is, we were interested in the way that peoplewho were already interacting were arranged. Subsequent studies used experi-mental procedures in that Ss were asked to interact, and observations weremade as to how they arranged themselves. Schizophrenic patients are also usedas subjects. We hope to learn if their difficulties in communicating with othersare reflected in the way they place themselves. Will they sit too close to theother person or too far away? We are also interested in possible sex differencesin seating arrangements. In all studies, every effort was made to secure naturalconditions, i.e., to locate the experiments in settings where interaction was cus-tomary and where the subjects would feel comfortable.

EXPERIMENT ONE

METHOD

The first study was conducted in the staff dining hall of a lSOO-bed mentalhospital. Most staff members eat in this hall, hence it serves nurses, plumbers,accountants, secretaries, etc. The hall was 36 by 68 feet and contained thirteentables (36 by 72 inches). Each of these tables consisted of two 36 by 36-inchtables placed together and with eight chairs arranged around it. One of thesetables, along with the chairs (lettered for identification) is shown in Figure 1.Service in the dining hall was cafeteria style. The staff members secured theirplates at the front of the cafeteria, went in line, were served, and then sat atwhichever table they chose.

In this study we were interested in the chairs occupied by people who wereinteracting. That is, did communication occur between the head and footpositions (A-E in Figure 1), or between people seated alongside of one another

STUDIES IN PERSONAL SPACE 249

FIGURE 1

Table and Arrangement of Chairs Used in Experiment

(B-C or G-H)? This study was purely observational and, in one sense, ex-ploratory in that no specific hypotheses were formulated.

Two observers were used throughout the study and all observations tookplace during the noon meal. Both O's would secure their food and then selecta table at which to sit. On some days the O's sat at the front of the hall, some-times at the rear, etc. The O's would then look around, and if both agreedthat they had a clear view of another table, they would wait two minutes, andthen record five seconds of interactions.

The recording was done on mimeographed forms, similar to Figure 1. Eachwould independently record the verbal interactions that ensued during thefive-second period. If no interactions occurred, another table was selectedlater. Usually two records were made (by each O) at a meal, and the observa-tions were done over a two-month period. In all, a total of 50 observations wasmade by each 0 .

KESUI-TS

The 50 interaction records for each observer were summed to find the num-ber of conversations between people sitting in all possible combinations ofchairs (A-B, A-C, F-G, F-H, etc.). The totals for Observer One were correlatedwith the totals for the second observer. The Pearson r between the two oh-

250 SOCIOMETRY

TABLE 1

Interactions between People in Adjacent and Distant Chairs

Chance Results ofexpectancy, 50 observations,

per cent per cent

Adjacent chairs 39 73Distant chairs 61 27

servers was .91, which is significant beyond the .01 level. This showed that theobservations were reliable. In the rest of this paper we will use the averagefigures for the two observers.

The first distinction that can be made is between chairs next to one anotherand chairs at a distance. That is, communication is not only possible betweenpeople in neighboring chairs, but also between people whose chairs are sepa-rated by other chairs. These latter can be called "distant chairs" in contrastto the "near chairs." The results of the 50 observations, based on the averagesfor the two O's, are presented in Table 1. These data show that of the 67recorded interactions, there were only 18 interactions between people sittingin distant chairs. This differs markedly from the figures based on the permuta-tions and shows that the people interacted with those sitting in neighboringchairs. This is hardly unexpected, but it is necessary to recognize this trendfirst before making a separate analysis of the interactions between people inneighboring chairs.

If all permutations of possible interactions between neighboring chairs arecomputed, we should predict, on the basis of chance, 36 per cent of the inter-actions between people sitting side by side, 27 per cent between people sittingface to face, and 36 per cent between people sitting corner to comer. Both theexpected frequencies and the obtained frequencies from the 50 observationsare presented in Table 2.

The results in Table 2 show that the interactions between people seatedcorner to corner (i.e., A-B, A-H, E-D, E-F) exceeded chance expectancy while

Arrangement

Corner-to-cornerSide-by-sideFace-to-face

TABLE 2

Arrangements of People Interacting

No. of pairsin this arrangement

30165

Chanceexpectancy

18.518.S13.9

STUDIES IN PERSONAL SPACE 251

those between people seated face to face were less than chance expectancy. Atest of significance of these data yields a chi-square of 19.1, p < .01.

DISCUSSION

These results parallel those of Festinger, et al. (3) and James (7) in show-ing that communication tends to take place between neighbors. Our resultsalso showed that the corners of the table were the loci of most of the inter-actions.

We felt that these results made it necessary to determine if corner-to-cornerinteraction would predominate in other size tables, and also to see whetherpeople who desire to interact will choose to sit corner to corner when presentedwith other possibilities. That is, there is the possibility that our results werenot so much due to the corner positions sparking the interaction as to peoplewho wanted to talk seating themselves in the corner positions.

EXPERIMENT TWO

In this study we hoped to learn the way people arrange themselves when theydesire to interact. Our method correspondingly changed from natural observa-tion to active experimentation. Our goal was to ask groups of subjects to inter-act and then observe the ways in which they arranged themselves.

PROCEDURE

In order to conceal the purpose of the experiment from the subjects, weasked them to discuss particular proverbs and told them that they would beinterviewed as to the meaning of the proverbs. The subjects were ushered tothe door of the cafeteria and told:

This is a study of discussion groups and the way people discuss things. On the cardhere is a statement I would like you to discuss. Please go into the cafeteria, sit down, andbegin discussing it. After a while you will be interviewed as to what you discussed.

Various groups were used under different conditions. These will be describedin detail in the next section. All subjects were interviewed as to the meaningof the proverb. During the interview E recorded how the people were sitting.It was apparent that none of the subjects suspected the purpose of the experi-ment.

RESULTS

A. In this part of the study six pairs of subjects were asked to sit at the typeof table shown in Figure 1 and discuss the proverbs. The subjects were assortedemployees (secretaries, personnel workers, nurses, etc.) and the groups werehomogeneous as to sex. All six pairs arranged themselves at the corners (A-B,

252 SOCIOMETRY

A-H, E-D, or E-F). On the basis of permutations of all possible arrangements,this result is highly significant (p < .Ol).^

B. The same procedure was repeated in another cafeteria which containedtables 35 by 48 inches. There were four chairs around each table. The subjectsused were ten pairs of hospital employees, again a heterogenous group of hos-pital employees (accountants, nurses, recreation workers, etc.). The resultsshowed that eight of the pairs arranged themselves corner to corner, while twoarranged themselves face to face.

C. The experiment was repeated in the staff dining hall (used in Study Oneand Study Two, Part A), but only single tables (36 by 36 inches) were used.The subjects could either sit face to face or corner to corner. The subjects usedwere nine pairs of student nurses from a nearby general hospital. The resultsshowed that eight of the pairs arranged themselves corner to corner while onepair chose a face-to-face arrangement.

D. The experiment was repeated again, with groups of three subjects each.This time the topic discussed was "If you could help design a mental hospitalward, how many people would you prefer in a bedroom?" The subjects wereeleven groups of three persons each, all members of a mental health associationof a distant city. The results showed that nine of the groups chose the cornerpositions (H-A-B or D-E-F) while two groups chose other arrangements(H-B-DandC-D-E).

E. In view of the many discussions of the difficulties that schizophrenicpatients have in communicating with others, the procedure was repeated usingsome of our hospital patients. We hoped to learn if the schizophrenic patientswould avail themselves of the corner-to-corner arrangement chosen by theother subjects.

The subjects in this study were 26 pairs of schizophrenic patients in thishospital. Most had been used in other researches and had been reasonablycooperative. Nine of the S's were used twice, but the second time each waspaired with a new partner. There were 38 males and 5 females in the sampleand the average age of the group was 38.9 years.

As a control group, U pairs of nonschizophrenic mental patients in this hos-pital (depressives, alcoholics, psychopaths, etc.) were tested.

The subjects were taken in like-sex pairs to the entrance of the cafeteria,given cards containing a proverb, and asked to go inside, sit down, and discussit.

The results are presented in Table 3. It shows that the nonschizophrenicpatients resembled the normal group in that they used the corner positions

2 The significances are evaluated by two-tailed chi-square tests, except where the fre-quencies are too small. Thpn fishef's exact methods are employe^.

STUDIES IN PERSONAL SPACE 253

more than would be expected by chance (p < .05). The schizophrenics, how-ever, did not show any such trend. The greatest number chose to sit in "dis-tant" arrangements. None of the normal or nonschizophrenic groups used thistype of arrangement.

Observation of the session disclosed that very few of the schizophrenicstalked at all. Most had wandered aimlessly up the aisles and sat down any-where. They would face the floor or away from their partner. They seemedto lack both direction and interest.

TABLE i

Seating Pattems of Pairs of Patients

Opposite Comer Side Distant

Schizophrenics 8 4 4 10Nonschlzophrenics 3 7 1 0

EXPERIMENT THREE

METHOD

The studies that follow all employed a "decoy." This was a person who wasa confederate of E and who was already seated at a particular chair before thesubject entered the room. The subject would then be asked to go inside anddiscuss a topic with the other person (i.e., with the seated decoy). The decoywas instructed to sit passively and look down at the typewritten card until Sbegan to sit down. Then the decoy would look up, greet S, and begin discussingthe topic.

The normal Ss used in the study were a heterogeneous group of hospital em-ployees and visitors. Many of the sample were used in two or more procedures,e.g., both with a male decoy and with a female decoy. No S discerned the actualpurpose of the study. The seating usually took place automatically, i.e., with-out conscious decision. A few of the patients had to be urged to sit down, butthis was very rare in the normal group. Most of the decoys were members ofthe Research Department, although several other staff members were used.In almost all conditions, at least two decoys of each sex were used.

The study is divided into three parts, depending upon the seat occupied bythe decoy.

Par t i . Decoy at £8Part 2. Decoy a t ePart 3. Decoy at B—no corner chair (i.e.. Chair A removed)

^Tbe chairs will be identified according to the letters shown in Fig. 1.

254 SOCIOMETRY

The results are presented in terms of the relationship between the chairselected by S and the chair occupied by the decoy. If the decoy was at B, thesubject could sit alongside him (C), facing him (H), at the corner (A), or atone of the "distant" chairs {D, E, F, G). The schizophrenic Ss were all patientsat this hospital.

RESULTS

Part 1: Decoy at B

The results for this condition are summarized in Table 4. The data areclassified as to (a) the sex of the decoy; (b) the sex of the S; and (c) whetherthe Ss were normal or schizophrenic.

It can be seen that there are some striking differences in the positions occu-pied by the Ss. With the female decoy the female (normal) Ss tended to sit inthe corner position. Males were more inclined to sit opposite both male andfemale decoys. The results from the female Ss with the female decoy differsignificantly at beyond the .05 level from the other three groups.

There is a marked difference (p < .01) between the normals and theschizophrenics in this condition. The schizophrenics sat alongside the decoyfar more than the normals and made almost no use of the corner chair (.,4).

TABLE 4

Seats Chosen When Decoy Sits at Seat B

Number of Ss choosing to sit:

Sex of Sex of Alongside Opposite Comer Opposite DistantGroup Ss decoy (C.D) (H) (A) side (F,G) corner (E) N

Male

Normals

Female

Male

Schizophrenics

Female

Male

Female

Male

Female

Male

Female

Male

Female

1

0

0

2

3

4

6

12

10

n

11

0

6

8

5

0

5

5

6

10

1

I

0

1

0

1

0

0

1

0

0

1

16

18

17

12

12

14

J2

14

STUDIES IN PERSONAL SPACE 255

Part 2: Decoy at C

The results when the decoy sat at C are summarized in Table 5. Again thefemales sit "closer" to the female decoy, although this time their preference isheavily weighted toward sitting alongside her. Only one of the normal malessat alongside a decoy. This difference is significant beyond the .01 level. The

Group

Normab

Schizophrenics

Sex ofSs

Male

Female

Male

Female

TABLE 5

Seats Chosen When Decoy Sits at Seat C

Sex ofdecoy

Male

Female

Male

Fcnaale

Male

Female

Male

Female

Number of Ss choosing to

AlongsideCB,D)

0

1

1

13

7

2

3

8

Opposite(G)

4

10

10

1

2

4

6

3

Corner(A,E)

4

8

4

4

6

5

1

5

sit:

Oppositeaide (F,H)

3

4

2

4

3

2

1

1

N

11

23

17

22

18

13

11

17

only significant difference between the normals and the schizophrenics in thistable is that seven of the male schizophrenics sat alongside the male decoywhile none of the male normals did.

Part 3: Decoy at B with No Comer Chair

The inclusion of this condition was dictated by our curiosity as to whetherface-to-face or side-by-side seating would predominate if corner seating wereprecluded. Thus the neighboring corner chair {A) was removed but otherwisethe procedure was identical to that of previous sessions. As only three of the52 schizophrenics had occupied the corner chair when it was present (in Part1), there seemed no point in assessing their reaction if it were absent. Henceonly normal Ss were used in this condition.

The results in Table 6 disclose the same trend for females to prefer sittingalongside the female decoy significantly more than males, or than either sexwith a male decoy.

256 SOCIOMETRY

Part 4: Some Further Studies

One surprising aspect of these results was that no trace was found of the"distant" seating pattern prevalent among the schizophrenic Ss of the pre-vious study. "Distant" seating refers to one person sitting in the chair that isseparated from his partner by another chair. In the previous study 10 of the26 pairs of schizophrenics had arranged themselves in this way. As many of

TABLE 6

Seats Chosen When Decoy Sits at Seat B (No Comer Chair)

Sex ofSs

Male

Female

Sex ofdecoy

Male

Female

Male

Female

Alongside(CD)

1

1

3

S

Number of Ss choosing to sit:

OppositeCH)

10

9

14

8

Oppositeside (F.G)

0

0

0

0

Distantcomer (E)

0

0

0

0

N

11

10

17

16

these Ss were also used in the present study, we had expected to find similarresults. Yet the data showed the Ss tended to sit too near rather than too farfrom the decoy.

However, the present procedure was very different from that of the previousstudy. Previously pairs of Ss had been escorted to the door of the cafeteria,given the topic, and asked to go inside, sit down, and discuss it. In this study,single Ss were brought to the door of the cafeteria, given the topic, and askedto go inside, sit down, and discuss it with a decoy who was already seated.The major differences were that the decoy was a staff member whereas thepartner had been a patient, and the S was now given a clear "target" (theseated decoy) toward which he could orient himself.

The procedure of the present study was thus repeated but with schizo-phrenic decoys. We were interested in learning whether the patients would sitdifferently vis-a-vis a schizophrenic patient than they had vis-a-vis a staffmember. Two separate experiments were carried out, one with a male decoyat seat B and another with a male decoy at seat C. To simplify matters onlymale decoys and male Ss were used. When the decoy sat a.i B, II Ss and 3separate decoys were used. When the decoy sat at C, 18 Ss and 4 decoyswere used.

STUDIES IN PERSONAL SPACE 257

The results are summarized in Table 7 and show that the status of the decoyas patient or nonpatient had very little influence on S's choice of a seat. Onethird of the patients sat side by side (which was very rare among the normalmales) and no trend toward "distant" seating was found.

However, unlike the previous procedure these Ss had now possessed a clear"target" or goal (the seated decoy) toward which they could orient themselves.Previously neither S had known where his partner was going to sit nor whatwas expected of him. This may account for the aimless wandering and the"distant" seating of Study Two. Other writers (4) have commented on the

TABLE 7

Seating Patterns of Schizophrenic Ss with Schizophrenic Decoys

Decoy at BDecoy at C

Total

Alongside

64

10

s

Opposite

36

9

chose to sit:

Corner

15

6

Oppositeside

03

3

Distantcomer

1

1

N

1118

29

importance of a clearly structured situation for schizophrenic patients. Ifthey do not know what is expected of them, their performance deteriorates andany group structure dissolves. This seems to have been what occurred whenthe two schizophrenics were asked to discuss a topic together. When a decoywas present, the S could be reasonably confident that if anything were expectedof him, the decoy would let him know. No such guidepost existed when twoschizophrenics were together. As research on mental hospital wards has shown(10, 11), spontaneous conversations between chronic schizophrenic patientsare very infrequent occurrences. Many of these patients will, however, talkto nurses if they are approached.

DISCUSSION

The observations of interactions in Part One show that in small groups thereis no simple relationship between distance and communication. Subjects whowere sitting side by side were physically closer to one another but interactedless than subjects sitting corner to corner. However, the trend in all the datais that people sitting in neighboring chairs (regardless of their positions) willbe more likely to interact than people sitting in distant chairs. Principlesgoverning spatial arrangements in small groups must, therefore, take intoaccount both the distance between people and their positions vis-a-vis eachother. Steinzor's (12) hypothesis that interaction will be more likely between

258 SOCIOMETRY

people who can see one another is supported by these data but must be qual-ified in terms of the angle between the participants. The subjects in thesestudies did not show any preference for face-to-face seating, where cornerseating was possible.

There are several conclusions to be drawn from the sessions in which thedecoy was used. Females will sit closer to a female than to a male. This iscloser than males will sit to decoys of either sex. Of the normal Ss only thefemales chose to sit alongside the decoy. The males overwhelmingly preferredthe chair opposite the decoy. This result parallels the observation that femalesin our culture will often be seen holding hands or kissing other females,whereas these behaviors are uncommon for males.

Obviously, there are cultural influences at work here. Perhaps if Fiji Island-ers were used as subjects they would prefer sitting on the tables insteadof on the chairs. Yet this does not diminish the relevance of this research forour own culture. The situation has parallels in animal research where ratsraised in one laboratory may be more gregarious or aggressive than rats raisedin another laboratory. Domesticated dogs or horses are undoubtedly very un-like their wild brethren, especially in regards to their "flight distance" (cf.Hediger), but this does not deny the importance of studying the habits ofdomestic animals. In fact, there are more practical reasons for studying thehabits and needs of captive and domestic animals than there are for studyingthe habits and needs of wild animals. The same holds true in the study ofhuman ecology. As long as man must live in a world of walls, furniture, doors,and fences, there is good reason to study how they influence his behavior.

The second conclusion from these data is that schizophrenic patients havean impaired concept of personal distance. This was especially evident whenthe schizophrenics of both sexes sat alongside the male decoy. This rarelyhappened in the normal group. One can only speculate how this affects therelations between the schizophrenic and his neighbors. He undoubtedly in-trudes on the personal space of those around him. This may cause the normalperson, whether a nurse or a relative, to draw away from him (i.e., to main-tain the customary personal distance). This may leave the schizophrenic feel-ing rejected by the withdrawal and the nurse feeling uncomfortable at theintrusion into her personal space.

It is possible to view this impairment of personal distance as a social arti-fact, but this time as a product of the mental hospital society. Usually amental patient has little or no privacy and very few things he can call hisown. The dayroom, dining hall, and especially the dormitory are often over-crowded. This may produce a situation analogous to that of the canaries inthe small cage reported by Allee ( I ) . When the area was crowded, no terri-tories were staked out, but upon transfer to a larger cage, the birds began to

STUDIES IN PERSONAL SPACE 259

select areas of their own. It would be interesting to observe in such a situationwhether prolonged imprisonment in a small cage would destroy or impair tbebirds' desire for individual territories.

It is curious that some of the most valuable research on the way that peopleor animals use space has occurred in situations of maximum constraint. Hed-iger's work with animals in zoos and circuses is exemplary. There is also agrowing literature on the ecology of mental hospital wards and geriatricscenters (9, 10, 11). Studies of human interaction in fields or courtyards arepractically nonexistent.

In closing it should be mentioned that our reason for conducting the studiesin cafeterias was not simply that other space was unavailable. Rather wehoped to achieve a setting where interaction would be relaxed and natural.We wanted to avoid, as much as possible, the use of a cold and impersonallaboratory, where our subjects might suspect hidden microphones and one-way mirrors. Following the methods of the animal biologists, we hoped for asetting where free interaction could occur. Our feeling is that the cafeteriaswere successful in getting the subjects to feel at ease. For example, the directorof a group of nurses that were subjects commented spontaneously, "The girlsare at home discussing things in a cafeteria."

SUMMARY

This was a series of studies of the ecology of small discussion groups. Thecentral question was the way that people will arrange themselves when theyare interacting. The setting of the study was a cafeteria in a large mental hos-pital, selected because it was a place where interaction could be free and"natural." The chairs around the rectangular tables were used as coordinatesfor locating the positions of the Ss. Ss could be described as sitting oppositeone another, alongside one another, at corner positions, or at some distancefrom each other. The study consisted of several related investigations.

1. Observations were made during the noon meal as to the position at thetable of the people who were talking. The results disclosed that people inneighboring chairs interacted more than people in distant chairs. Also thosein corner positions interacted more than people alongside one another or facingeach other.

2. Pairs of Ss were asked to enter the cafeteria and discuss various topics.Again a preference for the corner positions was seen. The same trend was evi-dent when groups of three Ss were used. Schizophrenic Ss made considerableuse of "distant" arrangements and would face away from their partners andrefuse to speak. However, nonschizophrenic mental patients resemble thenormal group in their preference for corner positions.

260 SOCIOMETRY

3. In this study, a "decoy" was used. This was a person who was alreadyseated at a table before S entered the room. The S was then asked to walkover and sit down and discuss a topic with him. Decoys and Ss of both sexeswere used in various combinations (i.e., male Ss with female decoys, male Sswith male decoys, etc.). The results disclosed that females would sit "closer"to female decoys than they would to male decoys; this was also "closer" thanmales would sit to decoys of either sex.

Schizophrenics Ss were also tested, and they appeared to have an impairedconcept of social distance. That is, male patients would sit alongside a maledecoy, female patients would sit alongside a male decoy, etc. This hardly everoccurred in the normal group.

Manuscript received: October 27, 1958Revised manuscript received: March 30,1959

Robert SommerSaskatchewan HospitalWeyburn, SaskatchewanCanada

REFERENCES

1. Allee, W. C, The Social Life of Animals. New York; Norton, 1938.2. Barker, R. G., and H. F, Wright, Midwest and Its Children. Evanston, HI.: Row,

Peterson, 1957.3. Festinger, L., S. Schachter, and K. Back, Social Pressures in Informal Groups. New

York: Harper, 1950.4. Goodrich, D. W,, L. Mazer, and B. Cline, "Fostering the Involvement of the Psychi-

atric Patient in Group Activities, Psychiatry, 1958, 21, 259-̂ 268.5. Hediger, H., Studies of the Psychology and Behavior of Animals in Zoos and Circuses.

New York: Criterion. 1955.6. Howard, E., Territory in Bird Life. London: John Murray, 1920.7. James, J., "Some Elements in a Theory of Small Groups," Research Studies of the

State College of Washington, 1950, 18, 150-151.8. Katz, David, Animals and Men. New York: Longmans, Green. 1937.9. Pace, R. E., "Situational Therapy," Journal of Personality, 1957, 25, 578-588.

10. Sommer, R., "Social Interaction on a Geriatrics Ward," International Journal of SocialPsychiatry, 1958, 4, 128-133.

11. Spohn, H. E,, et al., "Social Psychological Research Approaches in the Study of Schizo-phrenia." Paper presented at APA meeting, Washington, D. C, 1958.

12. Steinzor, B., "The Spatial Factor in Face-to-Face Discussion Groups," in Small Groups,ed. by P, Hare, E. Borgatta, and R. F. Bales. New York: Knopf, 195S.

13. Stern, W., General Psychology, translated by H. D. Spoerl. New York: Macmillan, 1938.14. Thra.sher, F. M., The Gang. Chicago: University of Chicago Press, 1927.15. Von Uexkull, J., "A Stroll through the Worlds of Animals and Men, in Instinctive Be-

havior, Claire Schiller, ed. New York: International Universities Press, I9S7.16. Whyte, W, F., Street Corner Society. Chicago: University of Chicago Press, 1943.17. Whyte, W. H., The Organization Man. New York: Simon and Schuster, 1956.